Hydrobiologia (2014) 727:1–18 DOI 10.1007/s10750-013-1779-z

PRIMARY RESEARCH PAPER

Macroinvertebrate colonization dynamics on artificial substrates along an algal resource gradient

A. Braccia • S. L. Eggert • N. King

Received: 13 August 2013 / Revised: 1 December 2013 / Accepted: 2 December 2013 / Published online: 14 December 2013 Ó Springer Science+Business Media Dordrecht 2013

Abstract Riparian canopy removal and land use also increased according to the gradient. Correlation may introduce multiple stressors that can alter food results indicated that chlorophyll a biomass, rather and habitat for stream organisms, but the influence of than time or temperature, was better related to the these alterations on macroinvertebrate colonization abundance and biomass of most primary consumers on dynamics is less well known. A field study involving substrates. These results suggest that the combined the simultaneous placement and removal of artificial effects of elevated temperatures and nutrients can substrates was performed to examine how macroin- result in relatively rapid algal accrual that may alter vertebrate colonization rates might vary with algal the colonization and establishment of macroinverte- accumulation within a perennial stream segment in brate communities in streams subjected to gradients of eastern Ohio, USA. The study was conducted over a riparian disturbances. 60-day summer colonization period in three reaches that were selected to represent an algal resource Keywords Functional feeding groups Á gradient according to canopy cover and agricultural Chlorophyll a Á Periphyton Á Nutrients Á nutrient sources in the riparian corridor. Total nitro- Agriculture Á Riparian canopy gen, water temperatures, and mean algal biomass from substrates increased along the resource gradient rep- resented by the study sites. Total macroinvertebrate Introduction biomass and the abundance and biomass of scrapers Identifying factors that influence the colonization dynamics of stream macroinvertebrates can improve Handling editor: Sonja Stendera our understanding of stream recovery following A. Braccia (&) natural and anthropogenic disturbance and help Department of Biological Sciences, Eastern Kentucky explain temporal variability in community structure University, Richmond, KY 40475, USA that is commonly observed in benthic macroinverte- e-mail: [email protected] brate studies. Since the degree and extent of distur- S. L. Eggert bance determines the size of habitat patches available USDA Forest Service, Northern Research Station, Grand for colonization, studies of macroinvertebrate coloni- Rapids, MN 55744, USA zation are conducted over large (e.g., rivers, streams; Williams & Hynes, 1977; Gray & Fisher, 1981; N. King Biology Department, Bucknell University, Lewisburg, Minshall et al., 1983a; Malmqvist et al., 1991; Milner PA 17837, USA et al., 2000) and small (e.g., stones, patches; Wise & 123 2 Hydrobiologia (2014) 727:1–18

Molles, 1979; Shaw & Minshall, 1980; Peckarsky, 1980; Williams, 1980; Robinson et al., 1993; Maton- 1986; Robinson et al., 1990) spatial scales. The ickin et al., 2001). Simultaneous measures of basal simultaneous placement and removal of artificial food resources (e.g., algae and detritus) have not substrates is an approach that has been widely used always been reported from colonization studies, so it is to study macroinvertebrate colonization of benthic difficult to generalize about how the accumulation of substrate patches following small-scale disturbance, these resources influence macroinvertebrate commu- such as flow-generated scour or tumbling of rocks nity colonization. However, continuous increases in (Ciborowski & Clifford, 1984; Lake & Doeg, 1985; macroinvertebrate densities and richness have been Robinson & Minshall, 1986; Downes & Lake, 1991; attributed to changes in habitat and the accumulation Rutherford, 1995; Baer et al., 2001; Miyake et al., of food resources on substrates over colonization 2003). intervals (Meier et al., 1979; Shaw & Minshall, 1980; The succession of taxa and trophic groups onto Baer et al., 2001), and more rapid colonization of benthic habitats has been well described from macr- substrates by macroinvertebrates has been reported oinvertebrate colonization studies. In general, highly from studies that used substrates with natural or mobile, opportunistic taxa within the collector-filter- artificial periphyton layers (Robinson et al., 1990; ing or -gathering functional groups will colonize first Baer et al., 2001; Miyake et al., 2003). Further, and then increases in the abundance of scrapers, Robinson et al. (1990) suggested that variable stabil- predators, and shredders can occur (see review by ization times for macroinvertebrate densities from Mackay, 1992). Colonization models predict that early small-scale colonization studies may be partly dynamics are primarily a result of stochastic factors explained by the presence of a periphyton layer. Since and that initial colonization should quickly increase to algal resources may provide food and habitat for maximum abundance, biomass, or taxa richness, and certain macroinvertebrates, equilibrium is unlikely to then remain at an equilibrium level (Sheldon, 1984; develop if substrates are continuously accumulating Minshall & Petersen, 1985). Time to equilibrium (i.e., epilithic material (Mackay, 1992). stabilization time) is often equated with plateaus (or Benthic food resources can be altered in stream maxima) in macroinvertebrate densities, biomass, or reaches where native forest has been removed from taxa richness. Results from small-scale colonization riparian zones because increased solar radiation and studies have shown that initial colonization is rapid temperatures can enhance standing stocks of periph- (2–4 days), but individuals of different taxa and yton and primary production (Lowe et al., 1986; Hill trophic groups colonize at different rates, and time to et al., 1995; Quinn et al., 1997; Kiffney et al., 2003). equilibrium in densities and richness is variable and Riparian canopy removal also reduces forest litter generally ranges from 4 to 30 days (Rosenberg & inputs that support the detrital-based food webs of Resh, 1982; Mackay, 1992). In some studies, equilib- temperate streams, and results from field studies have rium in total macroinvertebrate densities or richness shown that timber harvest from riparian zones can on substrates was never reached and successional increase the abundance of macroinvertebrates and changes in individual taxa extended beyond stabiliza- alter the functional composition of macroinvertebrate tion times for total densities or richness (Meier et al., communities (Newbold et al., 1980; Hawkins et al., 1979; Peckarsky, 1986; Baer et al., 2001; Miyake 1982; Wallace & Gurtz, 1986). In addition, agricul- et al., 2003). tural land use within watersheds can introduce inor- Variation in stabilization times for macroinverte- ganic nutrients (Broussard & Turner, 2009; Hill et al., brate densities has been attributed to season, experi- 2011) that may interact with light and elevated mental design, and the densities, mobility, and life temperatures and further stimulate primary produc- histories of available colonists (Rosenberg & Resh, tion. Altered riparian zones and nutrient inputs have 1982; Robinson et al., 1993; Miyake et al., 2003). been identified as major stressors to benthic macroin- Studies from temperate streams have demonstrated vertebrate communities (Dodds & Welch, 2000; that colonization rates of macroinvertebrates are more Paulsen et al., 2008). Results from the analysis of rapid during summer and early fall when elevated large observational datasets have revealed that macr- stream temperatures increase macroinvertebrate activ- oinvertebrate-based indicators of biological integrity ity and more recruits are present (Shaw & Minshall, (e.g., EPT taxa, total taxa richness) are negatively 123 Hydrobiologia (2014) 727:1–18 3 correlated with nutrients in wadeable streams in some region receives an average of 96 cm of precipitation regions of the U.S. (Miltner & Rankin, 1998; Wang per year. Based on coverage from the 2006 National et al., 2007; Yuan, 2010). Understanding the relative Land Cover Database (Homer et al., 2007), the role of algal resource accrual on macroinvertebrate watershed is predominantly forested (49%) with colonization dynamics may reveal mechanisms that cultivated crops and pasture (41%) and low/medium influence the structure of benthic macroinvertebrate intensity development (9%). Land use in the riparian communities in agricultural streams. We are unaware corridor is variable and areas of agricultural and light of any small-scale colonization studies that have residential use are interspersed among patches of addressed the question of how macroinvertebrate continuous forest. colonization of artificial substrates might be influ- The ‘‘low’’ end of the expected resource gradient enced by the combined effects of reduced riparian has a forested riparian zone (Forested Site) and is canopy cover and agricultural land use. located at the Hiram College’s James H. Barrow Field Since primary productivity can be enhanced by Station. The field station contains 200 acres of beech- light, temperature and nutrients, and macroinverte- maple forest and no agricultural nutrient sources. In brate colonization dynamics can be influenced by algal 2006, the Forested Site was classified as Coldwater accrual on substrates, we hypothesized that macroin- Habitat for fish and macroinvertebrates and fully vertebrate community colonization dynamics would supported its aquatic life use designation with an vary within a stream segment according to canopy Invertebrate Community Index Score of 52 (Ohio cover and agricultural nutrient sources within the Environmental Protection Agency, 2008). The other riparian corridor. The objective of our study was to study sites that represented ‘‘medium’’ and ‘‘high’’ measure macroinvertebrate colonization rates on arti- portions of the expected resource gradient have ficial substrates along an a priori selected gradient of minimal canopy cover—dubbed open sites—and are stress (represented by three study sites) during a single located upstream (Open Site 1) and downstream (Open season. Artificial substrates in this study served to Site 2) of the Forested Site. Open Site 1 is 3.5 km mimic bare patches of benthic habitat that are created upstream from the Forested Site and has riparian from flow-generated disturbance. We expected that vegetation that consists of grasses and shrubs that are algal resources would accumulate on substrates in all occasionally mowed by the land owner. Open Site 2 is study sites, but that rates of algal accumulation would 2 km downstream from the Forested Site and has differ among sites according to available light, nutri- similar riparian vegetation to Open Site 1. Land use ents, and water temperature. In turn, we hypothesized immediately adjacent to the Forested Site and Open that the total densities and biomass of macroinverte- Site 1 did not change between 2006 and the time of this brates, as well as certain taxa and trophic groups (e.g., study (2009). However, trees within the riparian zone scrapers), would increase with algal accumulation on of Open Site 2 were removed during a timber harvest substrates. operation in the lower portion of the watershed in 2008. No agricultural nutrient sources are adjacent to the riparian zones of the Forested Site and Open Site 1. Methods The nearest nutrient sources are row crops at approx- imately 800 m upstream of Open Site 1 and 980 m Study sites upstream of the Forested Site, and stream segments immediately upstream of both sites have forested The study sites were three, non-consecutive 100-m riparian zones. Open Site 2 receives drainage from row reaches in a segment of Silver Creek, Portage County, crop agriculture that is immediately upstream and OH, USA. Silver Creek is a third order tributary within adjacent to the riparian zone. Further, cattle have the upper Mahoning River watershed of the Ohio access to Silver Creek at approximately 700 m River basin (Fig. 1). The watershed is located within upstream of Open Site 2. We expected that the study the Erie-Ontario Lake Plain ecoregion (Omernik, sites—ordered Forested Site, Open Site 1, and Open 1987), and substrates of the streambed are shale and Site 2—would represent an algal resource gradient as a sandstone mixed with sand and gravel from Missis- result of potential stressors associated with riparian sippian and Pennsylvanian age lithographies. The conditions and nutrient sources described above. 123 4 Hydrobiologia (2014) 727:1–18

Fig. 1 Study site locations in Silver Creek, Mahoning River watershed, Portage County, Ohio, USA

Artificial substrates and sample processing Epilithic material from one tile of each set was analyzed for periphyton biomass and chlorophyll An artificial substrate set consisted of 20, a. For periphyton biomass,‘ the tile area was scraped 5.1 9 5.1 cm unglazed ceramic tiles mounted on and the slurry was filtered onto a preweighed glass plastic racks and secured to the stream bottom with fiber filter (0.45 lm pore size) and dried at 60°C until a metal stakes. Substrate sets were randomly placed in constant weight was achieved. Filters were ignited at shallow riffles at each study site on 8 June 2009 and 500°C for 24 h, desiccated, and reweighed to obtain left in place for 60 days. From initial placement, tiles ash-free dry mass (AFDM). The remaining ‘ tile area were collected on day (d) 2, d4, d8, d16, d32, d48, and was scraped and filtered onto a 0.45 lm glass fiber d60. On each collection day, one tile from each set was filter. Chlorophyll a was extracted from filters with collected for macroinvertebrates by placing a 250 lm 90% acetone and analyzed with a spectrophotometer mesh net behind the tile to capture macroinvertebrates after correcting for pheophytin (Lorenzen, 1967). that might be dislodged during removal. Another tile The entire area of the second tile was scraped and was collected for analysis of epilithic material. All the contents preserved in 70% ethanol for macroin- tiles were secured in dark containers, placed on ice, vertebrate identification and enumeration. Macroin- and transported to the laboratory for analysis. vertebrates were hand-sorted from organic matter 123 Hydrobiologia (2014) 727:1–18 5 under a dissecting microscope, identified to the lowest significant (P \ 0.05). Slopes from significant regres- practical taxonomic level, enumerated, and measured sions were compared with Tukey’s multiple compar- to the nearest 0.5 mm in length in order to estimate ison. If the interaction from ANCOVA was not biomass from published length–weight regressions significant (site x time, P [ 0.05), then the interaction (Benke et al., 1999). All insect taxa were identified to term was removed and ANOVA was performed on the genus, except Chironomidae, which were identified as main site effects (Engqvist, 2005). We also performed Tanypodinae or non-Tanypodinae. Non-insect taxa simple and partial Pearson correlations with combined were identified to phylum (Nematoda), class (Oligo- data from all study sites to examine the relationships chaeta, Turbellaria), subclass (Acari), and family between all taxa and trophic groups, and to explore the (Ancylidae). Taxa were then assigned to trophic relative influence of time, temperature, and chloro- groups according to information in Thorp & Covich phyll a on taxa and functional feeding group abun- (2009) and Merritt et al. (2008). dance and biomass. All response variables were log (x ? 1) transformed to meet the assumptions of normality prior to analysis. Analyses were performed Channel habitat and water chemistry in SASÒ v. 9.3 (SAS, Inc., Cary, NC, USA). Non-metric multidimensional scaling (NMS) was Channel width, water depth, and visual estimates of used to identify potential differences and gradients in bed substrate composition were measured at 11 macroinvertebrate community composition on tiles equally spaced transects within each site during within and among sites throughout the colonization summer 2009. Water temperature was continuously period. The NMS was conducted using PC-ORD v. 6.0 recorded at 15-min intervals with HOBO PendantÒ software (McCune & Mefford, 2011). Taxa that temperature loggers (model UA-001-64; Onset Com- occurred in less than 2% of samples were not included puter Corp., Bourne, MA, USA). Water velocity and in the analysis. We used the Bray–Curtis similarity depth were measured at artificial substrate locations on coefficient on the mean macroinvertebrate abundance five occasions during the colonization interval. Con- from tiles on each day of the colonization period in the ductivity, dissolved oxygen, and pH were determined main matrix and the data were computed with 250 from in situ readings when tiles were collected. Grab iterations, 50 real runs, and 50 randomized runs. To samples for nutrient analyses were collected on 27 and improve interpretation of the NMS, successional 31 July 2009 and 7 August 2009. Total phosphorus vectors were added to the ordination diagrams to (TP) and nitrogen (TN) were determined with a Lachat explore differences in compositional changes over Quik-Chem 8000 (Hach Co., Loveland, CO, USA) time at each site. Time (day during the colonization using in-line persulfate digestion and colorimetric period), degree days, and algal resources from tiles methods recommended by Lachat (Quik-Chem were correlated with NMS axes to determine the Method 10-115-01-3-A for TP and Method 10-107- relative influence of these variables on macroinverte- 04-3-P for TN). brate colonization patterns.

Data analysis Results Simple linear regression was used to determine if rates of algal resource accrual and macroinvertebrate den- Channel habitat and water chemistry sities and biomass increased linearly throughout the 60-day colonization period. Colonization rates were Relative to the Forested Site, the open sites had no or determined from slopes of significant regressions. minimal canopy cover, and greater amounts of fine Analysis of Covariance (ANCOVA) models were sediment (Table 1). Daily temperature ranges and performed to test for significant differences in colo- mean maximum temperatures at the open sites were nization rates among sites. Differences in rates of approximately 2–3X greater than those at the Forested increase/decrease were considered significantly dif- Site, and accumulated degree days (ADD) increased ferent if the interaction between the covariate (time) according to stressor gradient (Table 1). TP concen- and main factor (site) in the ANCOVA model was trations were similar among sites, but the maximum 123 6 Hydrobiologia (2014) 727:1–18

TN concentration detected at Open Site 1 and Open other sites (Table 3; Fig. 2). By d60, chlorophyll Site 2 was 2 and 7X greater than maximum concen- a biomass at Open Site 2 was approximately 3X the trations at the Forested Site (Table 1). amount at Open Site 1 and 12X times the amount at the Forested Site (Fig. 2). There was a significant linear Algal resource accrual relationship between chlorophyll a and mean maxi- mum TN concentrations from the study sites (SLR, ANCOVA indicated significant differences in accrual r2 = 0.99, P = 0.05) (Fig. 3). Among other variables rates of chlorophyll a and epilithic material on tiles in Table 1, chlorophyll a also was significantly among sites (Table 2). Since chlorophyll a and epi- correlated with ADD (r = 0.99, n = 3, P \ 0.05). lithic mass were highly correlated (r = 0.93, P \ 0.001), only chlorophyll a was used in further Macroinvertebrate richness, total density, analyses. Chlorophyll a increased linearly throughout and biomass the 60-day colonization period at all study sites and accrual rates at Open Site 2 were significantly more Twenty-one taxa, represented by four functional rapid than rates at the Forested Site, but rates at Open feeding groups—collector-gatherers, collector-filter- Site 1 were not significantly different from rates at the ers, scrapers, and predators—were collected from tiles during the colonization period. Nine taxa (non-Tany- podinae Chironomidae, Simulium, Baetis, Hydroptila, Cheumatopsyche, Antocha, Hydracarina, Tanypodi- Table 1 Habitat, physical–chemical, and nutrient variables nae, and Hemerodromia) comprised more than 90% of from each study site during the 60-day colonization period the total macroinvertebrate abundance at all study Forested Site Open Open sites. Throughout the colonization period, Open Site 1 Site 1 Site 2 had the greatest taxa richness with 18 taxa, while 13 Water velocity 0.3 ± 0.0 0.5 ± 0.1 0.5 ± 0.1 taxa were collected at both the Forested Site and Open (cm/s) Site 2. Tile depth (cm) 9.4 ± 0.9 6.4 ± 0.8 7.5 ± 0.9 ANCOVA indicated that most macroinvertebrate Channel Habitat responses had significantly different linear rates of Width (m) 6.8 ± 1.5 3.7 ± 0.9 5.9 ± 1.7 increase among sites (Table 2). Time to maximum Depth (m) 0.4 ± 0.3 0.2 ± 0.1 0.2 ± 0.2 richness did not show a clear increase along the % Fine sediment 13 ± 13 61 ± 29 44 ± 37 resource gradient, but taxa richness increased at a % Canopy cover 84 ± 14 1 ± 226± 38 faster rate at Open Site 2 (Table 3; Fig. 4a). Total Physical–chemical macroinvertebrate abundance had a weak linear Conductivity 229–466 322–480 271–480 increase during the 60-day colonization period at (lS/cm) Open Site 2 and ANCOVA indicated no difference in DO (mg/L) 7.7–8.6 7.1–10.1 7.7–11.3 rates among sites (Table 3; Fig. 4b). After the pH 7.75–8.09 7.72–8.06 7.74–8.47 site 9 time interaction term was removed, mean total Temperature (°C) macroinvertebrate abundance at the Forested Site was Daily mean 18.4 ± 1.4 18.6 ± 2.2 18.7 ± 2.3 significantly less than total abundance at open sites Daily max 19.4 ± 1.3 22.0 ± 2.0 22.0 ± 1.7 (Table 2, ANOVA: F2,60 = 27.12, P \ 0.001). By Daily range 2.0 ± 0.7 5.6 ± 1.9 5.3 ± 1.6 d60, total macroinvertebrate abundance at Open Site 2 ADD 1,083 1,088 1,097 was 4 and 8X the densities at Open Site 1 and the Nutrients (mg/L) Forested Site (Fig. 4b). Total macroinvertebrate bio- TP 0.04–0.04 0.03–0.05 0.02–0.05 mass increased linearly over time at the open sites and TN 0.15–0.27 0.23–0.47 0.37–1.89 the rate of increase was significantly faster at Open Site 2 (Table 3; Fig. 4c). Time to maximum biomass Values are means (±1SD) for habitat measures. Minimum and maximum values are reported for physical–chemical and increased along the resource gradient, and by d60, nutrient parameters. DO Dissolved oxygen; ADD total macroinvertebrate biomass at Open Site 2 was Accumulated degree days; TP Total phosphorous; TN Total 26X the amount at the Forested Site and 5X the nitrogen amount at Open Site 1 (Table 3; Fig. 4c). 123 Hydrobiologia (2014) 727:1–18 7

Table 2 Variables measured from tiles during the 60-day colonization period at each study site Variable Forested Site Open Site 1 Open Site 2 P value

Chlorophyll a (mg/m2) 1.9 ± 0.5 5.0 ± 1.6 13.7 ± 3.9 \0.0001 Epilithic mass (mg AFDM/cm2) 35.0 ± 8.1 251.2 ± 104.8 233.4 ± 75.9 \0.0001 Taxa richness (taxa/cm2) 3.1 ± 0.3 4.9 ± 0.4 5.2 ± 0.5 \0.0001 Macroinvertebrate abundance (ind/cm2) Total abundance 0.52 ± 0.06 2.26 ± 0.46 3.10 ± 0.37 0.0924 Collector-gatherers 0.28 ± 0.05 1.61 ± 0.49 1.18 ± 0.19 0.0003 Collector-filterers 0.21 ± 0.04 0.51 ± 0.17 1.51 ± 0.31 0.0330 Simulium 0.13 ± 0.04 0.43 ± 0.18 1.04 ± 0.33 \0.0001 Hydropsychidae 0.03 ± 0.01 0.07 ± 0.03 0.32 ± 0.11 \0.0001 Scrapers 0.00 ± 0.00 0.06 ± 0.03 0.36 ± 0.16 \0.0001 Predators 0.03 ± 0.01 0.08 ± 0.02 0.05 ± 0.02 \0.0001 Macroinvertebrate biomass (mg ADFM/cm2) Total biomass 0.034 ± 0.009 0.096 ± 0.021 0.363 ± 0.105 \0.0001 Collector-gatherers 0.001 ± 0.000 0.012 ± 0.004 0.036 ± 0.017 0.0010 Collector-filterers 0.025 ± 0.009 0.029 ± 0.009 0.292 ± 0.080 \0.0001 Simulium 0.001 ± 0.001 0.012 ± 0.005 0.031 ± 0.013 0.0047 Hydropsychidae 0.015 ± 0.008 0.017 ± 0.008 0.230 ± 0.080 \0.0001 Scrapers 0.000 ± 0.000 0.009 ± 0.004 0.030 ± 0.015 \0.0001 Predators 0.008 ± 0.005 0.047 ± 0.018 0.004 ± 0.003 \0.0001 Values are means (±1SE) (n = 21). P values are for the site 9 interaction from ANCOVA. P \ 0.05 indicates response variable had significantly different linear rates of increase among sites over the colonization period

Macroinvertebrate functional feeding groups and 8, Simulium (black flies) dominated total macro- invertebrate abundance (93–95%) on tiles at all study Collector-gatherers colonized tiles by d2 and were sites, and dominated total macroinvertebrate biomass abundant throughout the colonization period at all during that time at the open sites (Fig. 6a–f). Black fly sites (Fig. 5a). Collector-gatherer densities increased abundance decreased over time at all study sites, with linearly with time at the Forested Site and Open Site 2, a faster rate of decline at Open Site 2 (Table 4). but their rate of increase was more rapid at Open Site 2 Hydropsychid abundance and biomass increased lin- (Table 4). During the early colonization period early over time only at Open Site 2 (Table 4). (d2–8), chironomid midges were the only collector- Scraper taxa that colonized tiles between d16 and gatherers collected from tiles. By d16, the collector- 32 included Hydroptila and Baetis at all sites and gatherer Antocha occurred at all sites but was most Ancylidae at Open Site 1 (Fig. 6). Scraper abundance abundant on tiles at open sites. As a result of Antocha and biomass increased linearly at open sites, and rates abundance, collector-gatherer biomass increased lin- of increase were more rapid at Open Site 2 (Table 4; early with time and at similar rates at open sites Fig. 5e, f). The scraper contribution to total macroin- (Table 4; Fig. 5 b). Between d16 and 60, collector- vertebrate abundance and time to maximum scraper gatherer biomass at the Forested Site was dominated density and biomass clearly increased along the by midges (85%) and Antocha (60–75%) at the open resource gradient (Table 4; Fig. 6a–f). sites (Fig. 6a–f). Predators colonized tiles early (d2 or 4) at the Collector-filterers reached maximum densities by Forested Site and Open Site 1 but were not collected at d8, and then declined until the end of the colonization Open Site 2 until d16 (Fig. 5g, h). Early predators period at all study sites (Fig. 5c). Collector-filterer (d2–8) included large-bodied Perlesta, which contrib- biomass increased linearly during the colonization uted 26 and 64% to total macroinvertebrate biomass at period at Open Site 2 (Table 4; Fig. 5d). Between d2 Open Site 1 and the Forested Site (Fig. 6b, d). 123 8 Hydrobiologia (2014) 727:1–18

Table 3 Linear regression results for response variables from tiles and time (60-day colonization period) along the resource gradient (n = 21 for each study site) Adj P Colonization Days to r2 rate max

Chlorophyll a (mg/m2) Forested 0.64 \0.0001 0.0132A 32 Open 1 0.88 \0.0001 0.0210AB 60 Open 2 0.60 \0.0001 0.0250B 60 Taxa richness (taxa/cm2) Forested 0.19 0.0285 0.0038A 16 Open 1 0.25 0.0124 0.0033A 46 Open 2 0.67 \0.0001 0.0055B 32 Abundance (ind/cm2) Forested NS 46 Open 1 NS 46 Fig. 2 Chlorophyll a on tiles over the 60-day colonization period at each study site (n = 3 tiles). Values are mean ± 1SE. Open 2 0.17 0.0367 0.0038 60 Lines indicate significant linear increase over the 60 days (black Biomass line = Forested Site; gray line = Open Site 1; dashed line = (mg AFDM/cm2) Open Site 2) Forested NS 16 Open 1 0.44 0.0006 0.0012A 46 25 Open 2 0.60 \0.0001 0.0050B 60 20

NS indicates no significant linear relationship with time. Rates ) 2 (unit day-1) were determined from the slope of the linear regression line. Different letters indicate significant differences 15 (mg/m in rates following Tukey’s post hoc tests. Days to max is the a day when maximum values were detected 10 Hemerodromia, Hydracarina, and Tanypodinae Chlorophyll Chlorophyll occurred at all sites, Atherix and Chrysops occurred 5 Chlor a = -0.907+(16.315 x TN) 2 only at open sites, and Nigronia occurred only at the r = 0.99, p ≤ 0.05

Forested Site. Predator abundance increased linearly 0 at all sites, but the rates of increase did not correspond 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Total Nitrogen (mg/L) to the resource gradient (Table 4; Fig. 5g, h). How- ever, time to maximum predator density decreased Fig. 3 Relationship between mean ± 1SE total nitrogen con- along the gradient (Table 4). Late-colonizing preda- centrations in stream water (n = 3 dates) and chlorophyll tors at Open Site 1 included large Atherix, which a standing crop on tiles (n = 7 dates) during the colonization period at the study sites contributed to the linear increase in predator biomass at Open Site 1 (Table 4; Figs. 5h, 6c, d). chlorophyll a biomass than with time and ADD, and When data from all sites were combined, no the relationships remained significant when time and negative correlations between taxa or trophic groups ADD were held constant (Table 5). were detected. Most macroinvertebrate responses had similar correlations with time and ADD, since these Macroinvertebrate community composition variables were highly correlated (r = 0.99, n = 21, P \ 0.0001). With the exception of black flies, NMS produced a two-dimensional solution with a final abundance and biomass of most macroinvertebrate stress value of 9.25. NMS axis 1 represented 55% of primary consumers had stronger relationships with the variation in the community data and NMS axis 2 123 Hydrobiologia (2014) 727:1–18 9

Table 4 Linear regression results for response variables from tiles and time (60-day colonization period) along the resource gradient (n = 21 for each study site) Abundance Biomass Adj r2 P Colonization Days Adj r2 P Colonization Days rate to max rate to max

Collector-gatherers Forested 0.40 0.0013 0.0022A 46 NS 46 Open 1 NS 46 0.26 0.0105 0.0002A 46 Open 2 0.47 0.0004 0.0060B 16 0.20 0.0241 0.0007A 46 Collector-filterers Forested NS 46 NS 46 Open 1 NS 46 NS 46 Open 2 NS 60 0.58 \0.0001 0.0034 60 Simulium Forested 0.22 0.0184 -0.0006A 8NS8 Open 1 0.29 0.0074 -0.0044A 8NS8 Open 2 0.37 0.0021 -0.0073B 8 0.17 0.0373 -0.0005 8 Hydropsychidae Forested NS 46 NS 46 Open 1 NS 46 NS 46 Open 2 0.57 \0.0001 0.0050 60 0.67 \0.0001 0.0043 60 Scrapers Forested NS 16 NS 32 Open 1 0.18 0.0332 0.0010A 46 0.16 0.0401 0.0002A 46 Open 2 0.51 0.0002 0.0059B 60 0.42 0.0008 0.0009B 60 Predators Forested 0.33 0.0040 0.0006A 60 NS 8 Open 1 0.62 \0.0001 0.0013B 46 0.40 0.0012 0.0010 60 Open 2 0.49 0.0003 0.0010AB 32 NS 60 NS indicates no significant linear relationship with time. Rates (unit day-1) were determined from the slope of the linear regression line. Different letters indicate significant differences in rates following Tukey’s post hoc tests. Days to max is the day when maximum values were detected

represented an additional 30% (Fig. 7). The coloniz- correlated with ADD (r = 0.45) and time (r = 0.29) ing fauna from the Forested Site and Open Site 2 (Fig. 7a). Within each site, the greatest degree of separated considerably from each other in ordination change in community composition occurred during space, while Open Site 1 was more variable and the early colonization period (d2–16). Reversed intermediate to the other sites (Fig. 7a). Collector- trajectories with axis 2 on d8 at all sites resulted from gathering midges had the strongest positive correla- maximum black fly abundance (Fig. 7b). The direc- tion with axis 1 (r = 0.80). Eight other taxa also had tion of successional vectors along each axis indicates positive axis 1 correlations (r = 0.43–0.60). The the relative influence of chlorophyll a, ADD, and time predator Perlesta had a negative correlation (r = on macroinvertebrate colonization at each site -0.45) with axis 1 and Simulium had a strong negative (Fig. 7a, b). The chlorophyll a influence at Open Site correlation (r =-0.87) with axis 2. NMS axis 1 2 became especially apparent on d32 when the scores were positively correlated with chlorophyll successional vector reversed its trajectory along axis a (r = 0.51) while axis 2 scores were positively 2 and moved forward along axis 1 (Fig. 7a, b).

123 10 Hydrobiologia (2014) 727:1–18

correlation with algal accrual and clearly increased along the gradient. Most other macroinvertebrate metrics showed no clear increase along the gradient (e.g., total macroinvertebrate abundance, collector- gatherer abundance/biomass, predator abundance), and some metrics (e.g., predator biomass, abundance and biomass of hydropsychids) had relatively rapid rates of increase at only one site, which likely influenced the colonization rates of other taxa at those sites. Total biomass, rather than abundance, better reflected the resource gradient since biomass was not influenced by high densities of small-bodied black flies and chironomids. Relatively high densities of black flies during early colonization and fluctuating midge densities at the open sites resulted in relatively slow or no increase in total macroinvertebrate abun- dance during the colonization period. Black flies are opportunistic taxa that prefer clean substrates for attachment (Chutter, 1968; Hemphill & Cooper, 1983). Early colonization by black flies—followed by their quick declines—has been attributed to the development of unfavorable conditions as potential competitors (e.g., hydropsychids) and epilithic layers increase (Hemphill & Cooper, 1983; Ciborowski & Clifford, 1984; Downes & Lake, 1991). In our study, pupal cases were absent and we found no correlations between black flies and both chlorophyll a and other taxa. These findings suggest that black fly declines between d8 and 16 probably resulted from the development of unfavorable substrate conditions, rather than emergence or negative interactions with other taxa. Lack of continuous increases in macroinvertebrate biomass over the colonization period at the Forested Fig. 4 Macroinvertebrate taxa richness (a), abundance (b), and Site resulted from low densities of primary consumers. biomass (c) from tiles at each study site (n = 3 tiles) during the Food webs in temperate forested stream reaches are 60-day colonization period. Values are means ± 1SE. Lines detrital based and supported by inputs of forest litter indicate significant linear increase over the 60 days (black from the riparian zone (Fisher & Likens, 1973; line = Forested Site; gray line = Open Site 1; dashed line = Open Site 2) Minshall et al., 1983b; Wallace et al., 1999). Many taxa in forested stream reaches are adapted for feeding on leaf litter and detritus that accumulates in the Discussion interstitial spaces of stream beds and debris dams (Wallace & Webster, 1996). Tiles at the Forested Site Accumulated temperatures, TN concentrations, and probably lacked preferred food for taxa that are standing stocks of chlorophyll a biomass on substrates adapted for feeding on terrestrial litter inputs rather increased along the algal resource gradient as we than periphyton. Furthermore, macroinvertebrate col- expected. Total macroinvertebrate biomass and the onization rates are influenced by the densities, mobil- abundance and biomass of scrapers had high ity, and composition of the pool of available 123 Hydrobiologia (2014) 727:1–18 11

Fig. 5 Macroinvertebrate functional feeding group abundance (a, c, e, g) and biomass (b, d, f, g) from tiles at each study (n = 3 tiles) during the 60-day colonization interval. Values are means ± 1SE. Lines indicate significant linear increase over 60 days (black line = Forested Site; gray line = Open Site 1; dashed line = Open Site 2)

colonizers (Shaw & Minshall, 1980; Benson & densities in riffles, as a result of decreased primary Pearson, 1987; Robinson et al., 1993). Several studies production in the reach, might also explain the have shown that macroinvertebrate densities and relatively slower colonization by macroinvertebrates biomass (or both) can be greater in open canopy at the Forested Site. stream sections presumably as a result of altered food The biomass response to the gradient became resources from increased primary production (Haw- apparent on d32 when total macroinvertebrate bio- kins et al., 1982; Behmer & Hawkins, 1986; Dudgeon mass at Open Site 2 differed from the other sites due to & Chan, 1992; Quinn et al., 1997). Decreased increases in hydropsychid densities since d16. 123 12 Hydrobiologia (2014) 727:1–18

Table 5 Results from simple and partial Pearson correlations between mean macroinvertebrate abundance (Abd) and biomass (Bio), time (day), degree days, and mean chlorophyll a biomass (mg/m2) from each sampling date for all sites combined (n = 21) Simple correlations Partial correlations Time Degree days Chlorophyll a Chlorophyll a rPrPrP r P

Abundance 0.32 NS 0.34 NS 0.51 0.0179 0.55 0.0152 Biomass 0.53 0.0141 0.53 0.0127 0.70 0.0004 0.59 0.0075 Collector-gatherers Abd 0.63 0.0023 0.64 0.0018 0.73 0.0002 0.55 0.0151 Bio 0.57 0.0076 0.56 0.0081 0.77 \0.0001 0.66 0.0021 Collector-filterers Abd -0.34 NS -0.33 NS -0.14 NS 0.36 NS Bio 0.33 NS 0.34 NS 0.51 0.0187 0.49 0.0329 Simulium Abd -0.70 0.0004 -0.68 0.0006 -0.58 0.0061 0.11 NS Bio -0.53 0.0134 -0.52 0.0162 -0.46 0.0345 0.00 NS Hydropsychidae Abd 0.48 0.0276 0.49 0.0254 0.66 0.0012 0.56 0.0119 Bio 0.53 0.0129 0.54 0.0119 0.68 0.0007 0.53 0.0200 Scrapers Abd 0.62 0.0028 0.61 0.0031 0.73 0.0002 0.49 0.0329 Bio 0.56 0.0087 0.55 0.0100 0.71 0.0003 0.54 0.0171 Predators Abd 0.66 0.0010 0.65 0.0013 0.57 0.0071 -0.01 NS Bio 0.31 NS 0.30 NS 0.28 NS 0.04 NS Values are correlation coefficients (r) followed by P values. Partial correlations were determined by holding time and degree days constant

Collector-filtering hydropsychids usually increase on Rapid hydropsychid colonization rates at Open Site 2 substrates after the development of a periphyton layer could have also resulted from greater densities and because they prefer irregular surfaces for attachment production in riffle habitats. Elevated temperatures of their retreats (Hynes, 1975; Hemphill & Cooper, and improved food quality, such as algal-derived 1983; Mackay, 1992). Greater chlorophyll a standing seston, may increase the abundance and production of crops on d16 at Open Site 2 suggests that favorable collector-filterers (Wallace & Merritt, 1980). The habitat for hydropsychids developed earlier than at discrepancy between hydropsychid abundance and other sites. Hydropsychids can facilitate colonization biomass between d32 and 46, and the strong relation- of other taxa because their retreats enhance habitat, ship between hydropsychid biomass and time (rather substrate complexity, and food resources (Diamond, than abundance and time), indicates that an increase in 1986; Englund, 1993; O’Connor, 1993; Nakano et al., their biomass was possibly due to growth or selective 2005). Cardinale et al. (2001) showed that early replacement of small individuals by larger individuals. colonization of substrates by Ceratopsyche bronta Though hydropsychids can facilitate colonization of resulted in greater abundance and biomass of all certain taxa, they can also be defensive and territorial, macroinvertebrates after a 30-day colonization period. thus preventing the establishment of other taxa These attributes of hydropsychids—coupled with their (Englund & Olsson, 1990; Hemphill, 1991; Englund, high colonization rates—may account for the more 1993). Relatively large hydropsychids may have rapid increase in taxa richness noted at Open Site 2. excluded other large taxa, which might explain why

123 Hydrobiologia (2014) 727:1–18 13

Fig. 6 Macroinvertebrate relative abundance (a, c, e) and biomass (b, d, f) on tiles from study sites throughout the 60-day colonization period

most predators from tiles at Open Site 2 were correlation between scraper abundance and biomass relatively small (e.g., mites) compared to large suggests continuous colonization by individuals. The predators at other sites (e.g., Atherix and Nigronia). scraper response to algal accrual in this study is not Scrapers also contributed to biomass increases on surprising since results from field studies have shown tiles between d32 and 60 at the open sites. The high direct relationships between scrapers and algal 123 14 Hydrobiologia (2014) 727:1–18

Fig. 7 Non-metric a multidimensional scaling results. Circles represent Forested Site mean macroinvertebrate Open Site 1 densities (n = 3 tiles) on Open Site 2 each day of the colonization period. a Biplot overlay shows the correlation of time, degree days, and Degree days chlorophyll a with ordination axes 1 and 2— length of vectors reflects the Time Chlorophyll a magnitude of the correlation with each axis. b Time trajectories show how macroinvertebrate

community composition Axis 2 (30%) shifted along axes over the 60 day colonization period—numbers next to circles indicate day during the colonization period

Axis 1 (55%) b

60

60 46

32 32 46

4 4 16 16 16 2 2

32 46 60 8 Axis 2 (30%)

4 8

2

8 Axis 1 (55%)

123 Hydrobiologia (2014) 727:1–18 15 densities (Wallace & Gurtz, 1986; Richards & Min- indicated by maximum values on the last day of the shall, 1988; Dudgeon & Chan, 1992). Further, in colonization period. Maximum densities and biomass small-scale colonization studies, scraper densities on on d60 resulted from continuous increases in the substrates can be tightly coupled with their algal food abundance and biomass of scrapers and hydropsychids resources, and scraper declines have been attributed to through the last day of the colonization period at Open food depletion by their grazing activity (Lamberti & Site 2. Over a 65-day colonization interval in a Resh, 1983). Food depletion could explain the early forested stream, Baer et al. (2001) also found lack of decline in scraper densities at the Forested Site where stabilization in macroinvertebrate densities on sub- scraper abundance declined following decreases in strates, which they attributed to continuous increases chlorophyll a. However, food depletion is an unlikely in resource accumulation on artificial substrates. explanation for scraper declines on d46 at Open Site 1 Given that scraper densities can increase with algal since algal resources were still accruing through d60. resources on substrates (Feminella & Hawkins, 1995), The increase in macroinvertebrate biomass during the the clear increase in day to maximum scraper abun- mid- to late-colonization period at Open Site 1 also dance and biomass along the stressor gradient provides resulted from increased predator densities and then some evidence that stabilization times might have replacement of smaller predators by the large predator been delayed as a result of differences in algal accrual Atherix. On d60, several large Atherix and smaller rates among sites. The negative trend between day to predators were collected from every tile at this site. maximum predator abundance and the resource gra- Though we did not detect significant negative corre- dient suggests that predators may have also been lations between taxa, abundance of large predators on altered by algal accrual rates. Shorter stabilization d60 corresponding with declines in densities of all times for predator abundance at the open sites could primary consumers may simply be due to predation at have resulted from interactions with other taxa (such Open Site 1. Predators have been shown to influence as hydropsychids), more rapid rates of prey accumu- prey densities on substrate patches in streams (Lan- lation, or reduced predator drift rates. Results from caster et al., 1991). At Open Site 2, continuous experimental studies have shown that macroinverte- increases in algal resources and scraper colonization brate emigration rates (via drift) may be low in stream throughout the mid- to late-colonization period sug- areas where food resources are abundant (Richardson, gests that food resources were not limiting. 1991; Hinterleitner-Anderson et al., 1992; Siler et al., With the exception of predators, the timing and 2001). If predators drift less where algal resources and sequence of colonization was similar among sites and prey are abundant, then predators should have did not appear to be directly influenced by resource encountered tiles less frequently, which might explain accrual. The apparent absence of predators, especially their delayed arrival on tiles at Open Site 2. Perlesta, during the early colonization period at Open Results from correlation and ordination analyses Site 2 is puzzling since data from previous benthic suggest that the abundance and biomass of macroin- collections indicated its presence in riffles following vertebrate primary consumers on substrates mainly canopy removal (Braccia, unpublished data). Given increased with chlorophyll a biomass, though time and our small sample size, we are uncertain if Perlesta temperature were also important. The total macroin- absence resulted from a sampling artifact or factors vertebrate biomass response to the gradient was partly associated with more rapid resource accrual. Although driven by colonization of primary consumers that are we did not detect significant negative correlations favorably influenced by algal resources as food or between taxa in this study, the presence of Perlesta at habitat (e.g., scrapers and hydropsychids), which the Forested Site and Open Site 1 could have altered further suggests that algal accumulation was relatively colonization rates of other taxa during the early more important than time and temperature. The factors colonization period because predaceous stonefly lar- that influence benthic algal biomass in wadeable vae can reduce immigration rates of prey taxa on streams are complicated, and include light, tempera- substrate patches (Peckarsky & Dodson, 1980; Pec- ture, nutrients, discharge, and macroinvertebrate graz- karsky, 1985). ing (Dodds & Welch, 2000). Our study was not Total macroinvertebrate densities and biomass did designed to determine factors that limit benthic algae not appear to stabilize on tiles at Open Site 2, as in Silver Creek, so we are unable to clearly identify 123 16 Hydrobiologia (2014) 727:1–18 whether greater chlorophyll a biomass and accrual Acknowledgments This work was supported by the James H. rates resulted from elevated light, nutrients, temper- Barrow Field Station at Hiram College through a generous gift from the Paul and Maxine Frohring Foundation. We are ature, or a combination of these factors. However, especially grateful for the hospitality of a private landowner cattle access and nearby row crop agriculture corre- and assistance from students and staff at the field station. We sponded with elevated TN concentrations, greater thank Doris Nelson and John Larson of the USDA Forest standing stocks of algal biomass, and more rapid Service Northern Research Station for assistance with water chemistry analyses, Tom Koehnle and Matt Hils for assistance colonization of substrates by certain taxa (e.g., hyd- in the field, and Bob Butler and Daniel Douglas for their ropsychids) and trophic groups (e.g., scrapers) at Open comments on an earlier draft of this manuscript. Site 2. These results might suggest that TN contributed to increased primary productivity, which in turn resulted in increased rates of macroinvertebrate col- References onization. Recent experimental evidence also suggests that addition of nutrients can alter macroinvertebrate Baer, S. G., E. R. Siler, S. L. Eggert & J. B. Wallace, 2001. Colonization and production of macroinvertebrates on community responses to small-scale disturbance by artificial substrata: upstream–downstream responses to a changing the structure of primary consumer assem- leaf litter exclusion manipulation. Freshwater Biology 46: blages (Gafner & Robinson, 2007). 347–365. Land use within watersheds can introduce multiple Behmer, D. J. & C. P. Hawkins, 1986. Effects of overhead canopy on macroinvertebrate production in a Utah stream. stressors to stream ecosystems (Ormerod et al., 2010). Freshwater Biology 16: 287–300. Although findings from artificial substrate studies Benke, A. C., A. D. Huryn, L. A. Smock & J. B. Wallace, 1999. have revealed much about the colonization and Length–mass relationships for freshwater macroinverte- establishment of macroinvertebrate communities fol- brates in North America with particular reference to the southeastern United States. Journal of the North American lowing small-scale disturbance in streams, the influ- Benthological Society 18: 308–343. ence of stressors—such as elevated temperature and Benson, L. J. & R. G. Pearson, 1987. The role of drift and effect nutrients—on macroinvertebrate colonization dynam- of season on macroinvertebrate colonization of implanted ics have not been widely studied. Understanding how substrata in a tropical Australian stream. Freshwater Biology 18: 109–116. stressors alter the colonization and establishment of Broussard, W. & R. E. Turner, 2009. A century of changing macroinvertebrate communities on benthic habitats land-use and water-quality relationships in the continental could improve our understanding of the macroinver- US. Frontiers in Ecology and the Environment 7: 302–307. tebrate community structure in impaired streams and Cardinale, B. J., C. M. Smith & M. A. Palmer, 2001. The influence of initial colonization by hydropsychid caddisfly increase our chances for successful stream restoration. larvae on the development of stream invertebrate assem- This study provides a description of macroinvertebrate blages. Hydrobiologia 455: 19–27. community colonization in a perennial stream seg- Chutter, F. M., 1968. On the ecology of the fauna of stones in the ment with patches of open canopy and agricultural current in a South African river supporting a very large Simulium (Diptera). Journal of Applied Ecology 5: land use along its riparian corridor. Our results show 531–561. that continuous increases in algal resources on Ciborowski, J. J. H. & H. F. Clifford, 1984. Short-term coloni- substrates and elevated temperatures in surrounding zation patterns of lotic macroinvertebrates. Canadian stream reaches can result in more rapid colonization of Journal of Fisheries and Aquatic Sciences 41: 1626–1633. Diamond, J. M., 1986. Effects of larval retreats of the caddisfly substrates by certain taxa and trophic groups and Cheumatopsyche on macroinvertebrate colonization in increase the stabilization times for total macroinver- piedmont, USA streams. Oikos 47: 13–18. tebrate densities and biomass. These results lend Dodds, W. K. & E. B. Welch, 2000. Establishing nutrient cri- support to Robinson et al.’s (1990) suggestion that teria in streams. Journal of the North American Bentho- logical Society 19: 186–196. periphyton might account for highly variable stabil- Downes, B. J. & P. S. Lake, 1991. Different colonization pat- ization times in colonization studies and provides terns of two closely related stream insects (Austrosimulium evidence that algal accumulation and macroinverte- spp.) following disturbance. Freshwater Biology 26: brate colonization of substrates can vary according to 295–306. Dudgeon, D. & I. K. Chan, 1992. An experimental study of the canopy cover and adjacent land use within a stream influence of periphytic algae on invertebrate abundance in segment. a Hong Kong stream. Freshwater Biology 27: 53–63.

123 Hydrobiologia (2014) 727:1–18 17

Englund, G., 1993. Interactions in a lake outlet stream com- Lamberti, G. A. & V. H. Resh, 1983. Stream periphyton and munity: direct and indirect effects of net-spinning caddis insect herbivores: an experimental study of grazing by a larvae. Oikos 66: 431–438. caddisfly population. Ecology 64: 1124–1135. Englund, G. & T. Olsson, 1990. Fighting and assessment in the Lancaster, J., A. G. Hildrew & C. R. Townsend, 1991. Inver- net-spinning caddis larvae Arctopsyche ladogensis: a test tebrate predation on patchy and mobile prey in streams. of the sequential assessment game. Animal Behaviour 39: Journal of Animal Ecology 60: 625–641. 55–62. Lorenzen, C. J., 1967. Determination of chlorophyll and pheo- Engqvist, L., 2005. The mistreatment of covariate interaction pigments: spectrophotometric equations. Limnology and terms in linear model analyses of behavioural and evo- Oceanography 12: 343–346. lutionary ecology studies. Animal Behaviour 70: 967– Lowe, R. L., S. W. Golladay & J. R. Webster, 1986. Periphyton 971. response to nutrient manipulation in streams draining Feminella, J. W. & C. P. Hawkins, 1995. Interactions between clearcut and forested watersheds. Journal of the North stream herbivores and periphyton: a quantitative analysis American Benthological Society 5: 221–229. of past experiments. Journal of the North American Ben- Mackay, R. J., 1992. Colonization by lotic macroinvertebrates: a thological Society 14: 465–509. review of processes and patterns. Canadian Journal of Fisher, S. G. & G. E. Likens, 1973. Energy flow in Bear Brook, Fisheries and Aquatic Sciences 49: 617–628. New Hampshire: an integrative approach to stream eco- Malmqvist, B., S. Rundle, C. Bronmark & A. Erlandsson, 1991. system metabolism. Ecological Monographs 43: 421–439. Invertebrate colonization of a new, man-made stream in Gafner, K. & C. T. Robinson, 2007. Nutrient enrichment southern Sweden. Freshwater Biology 26: 307–324. influences the responses of stream macroinvertebrates to Matonickin, R., I. Habdija & B. Primc-Habdija, 2001. The disturbance. Journal of the North American Benthological effects of season and food availability on macroinverte- Society 26: 92–102. brate colonization in a woodland stream. Archiv fur Hy- Gray, L. J. & S. G. Fisher, 1981. Postflood recolonization drobiologie 153: 55–74. pathways of macroinvertebrates in a lowland Sonoran McCune, B. & M. J. Mefford, 2011. PC-ORD. Multivariate Desert stream. American Midland Naturalist 106: 249–257. Analysis of Ecological Data. Version 6. MjM Software, Hawkins, C. P., M. L. Murphy & N. H. Anderson, 1982. Effects Gleneden Beach, Oregon, USA. of canopy, substrate composition, and gradient on the Meier, P. G., D. L. Penrose & L. Polak, 1979. The rate of col- structure of macroinvertebrate communities in Cascade onization by macro-invertebrates on artificial substrate Range streams of Oregon. Ecology 63: 1840–1856. samplers. Freshwater Biology 9: 381–392. Hemphill, N., 1991. Disturbance and variation in competition Merritt, R. W., K. W. Cummins & M. B. Berg, 2008. An between two stream insects. Ecology 72: 864–872. Introduction to the Aquatic Invertebrates of North Amer- Hemphill, N. & S. D. Cooper, 1983. The effect of physical ica. Kendall Hunt, Dubuque. disturbance on the relative abundances of two filter-feeding Milner,A.M.,E.E.Knudsen,C.Soiseth,A.L.Robertson,D. insects in a small stream. Oecologia 58: 378–382. Schell, I. T. Phillips & K. Magnusson, 2000. Colonization and Hill, W. R., M. G. Ryon & E. M. Schilling, 1995. Light limi- development of stream communities across a 200-year gra- tation in a stream ecosystem: responses by primary pro- dient in Glacier Bay National Park, Alaska, USA. Canadian ducers and consumers. Ecology 76: 1297–1309. Journal of Fisheries and Aquatic Sciences 57: 2319–2335. Hill, B. H., D. W. Bolgrien, A. T. Herlihy, T. M. Jicha & T. Miltner, R. J. & E. T. Rankin, 1998. Primary nutrients and the R. Angradi, 2011. A synoptic survey of nitrogen and biotic integrity of rivers and streams. Freshwater Biology phosphorus in tributary streams and great rivers of the 40: 145–158. upper Mississippi, Missouri, and Ohio River Basins. Minshall, G. W. & R. C. Petersen, 1985. Towards a theory of Water, Air and Soil Pollution 216: 605–619. macroinvertebrate community structure in stream ecosys- Hinterleitner-Anderson, D., A. E. Hershey & J. A. Schuldt, tems. Archiv fur Hydrobiologie 104: 49–64. 1992. The effects of river fertilization on mayfly (Baetis) Minshall, G. W., D. A. Andrews & C. Y. Manuel-Faler, 1983a. drift patterns and population density in an arctic river. Application of island biogeographic theory to streams: Hydrobiologia 240: 247–258. macroinvertebrate recolonization of the Teton River, Homer, C., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Larson, N. Idaho. In Barnes, J. R. & G. W. Minshall (eds), Stream Herold, A. McKerrow, J. N. VanDriel & J. Wickham, 2007. Ecology: Application and Testing of General Ecological Completion of the 2001 national land cover database for Theory. Plenum, New York. the conterminous Unites States. Photogrammetric Engi- Minshall, G. W., R. C. Petersen, K. W. Cummins, T. L. Bott, J. neering and Remote Sensing 73: 337–341. R. Sedell, C. E. Cushing & R. L. Vannote, 1983b. Interb- Hynes, J. D., 1975. Annual cycles of macro-invertebrates of a iome comparison of stream ecosystem dynamics. Ecolog- river in southern Ghana. Freshwater Biology 5: 71–83. ical Monographs 53: 1–25. Kiffney, P. M., J. S. Richardson & J. P. Bull, 2003. Responses of Miyake, Y., H. Tsutomu, N. Kuhara & S. Nakano, 2003. Suc- periphyton and insects to experimental manipulation of cession in a stream invertebrate community: a transition in riparian buffer width along forest streams. Journal of species dominance through colonization. Ecological Applied Ecology 40: 1060–1076. Research 18: 493–501. Lake, P. S. & T. J. Doeg, 1985. Macroinvertebrate colonization Nakano, D., Y. Masamichi & T. Okino, 2005. Ecosystem of stones in two upland southern Australian streams. engineering by larvae of net-spinning stream caddisflies Hydrobiologia 126: 199–211. creates a habitat on the upper surface of stones for mayfly

123 18 Hydrobiologia (2014) 727:1–18

nymphs with a low resistance to flows. Freshwater Biology an Idaho stream. Journal of the North American Bentho- 50: 1492–1498. logical Society 9: 240–248. Newbold, J. D., D. C. Erman & K. B. Roby, 1980. Effects of Robinson, C. T., G. W. Minshall & L. Van Every, 1993. Sea- logging on macroinvertebrates in streams with and without sonal trends and colonization patterns of macroinvertebrate buffer strips. Canadian Journal of Fisheries and Aquatic assemblages in two streams with contrasting flow regimes. Sciences 37: 1076–1085. Great Basin Naturalist 53: 321–331. O’Connor, N. A., 1993. Resource enhancement of grazing Rosenberg, D. M. & V. H. Resh, 1982. The use of artificial mayfly nymphs by retreat-building caddisfly larvae in a substrates in the study of freshwater benthic macroinver- sandbed stream. Australian Journal of Marine and Fresh- tebrates. In Cairns, J. (ed.), Artificial Substrates. Ann Arbor water Research 44: 353–362. Science, Ann Arbor, MI. Ohio Environmental Protection Agency (EPA), 2008. Biologi- Rutherford, J., 1995. Patterns of dispersion of aquatic insects cal and Water Quality Study of the Upper Mahoning River colonizing artificial substrates in a southern Ontario and Selected Tributaries 2006. EAS/2008-10-8. Ohio EPA, stream. Canadian Journal of Zoology 73: 458–468. Columbus, Ohio, USA. Shaw, D. W. & G. W. Minshall, 1980. Colonization of an Omernik, J. M., 1987. Ecoregions of the conterminous United introduced substrate by stream macroinvertebrates. Oikos States. Annals of the Association of American Geogra- 34: 259–271. phers 77: 118–125. Sheldon, A. L., 1984. Colonization dynamics of aquatic insects. Ormerod, S. J., M. Dobson, A. G. Hildrew & C. R. Townsend, In Resh, V. H. & D. M. Rosenberg (eds), The Ecology of 2010. Multiple stressors in freshwater ecosystems. Fresh- Aquatic Insects. Praeger Publishers, New York. water Biology 55: 1–4. Siler, E. R., J. B. Wallace & S. L. Eggert, 2001. Long-term Paulsen, S. G., A. Mayio, D. V. Peck, J. L. Stoddard, E. Tar- effects of resource limitation on stream invertebrate drift. quinio, S. M. Holdsworth, J. Van Sickle, L. L. Yuan, C. Canadian Journal of Fisheries and Aquatic Sciences 58: P. Hawkins, A. T. Herlihy, P. R. Kaufmann, M. T. Barbour, 1624–1637. D. P. Larsen & A. R. Olsen, 2008. Condition of stream Thorp, J. H. & A. P. Covich, 2009. Ecology and Classification of ecosystems in the US: an overview of the first national North American Freshwater Invertebrates. Academic assessment. Journal of the North American Benthological Press, New York. Society 27: 812–821. Wallace, J. B. & M. E. Gurtz, 1986. Response of Baetis mayflies Peckarsky, B. L., 1985. Do predaceous stoneflies and siltation (Ephemeroptera) to catchment logging. American Midland affect the structure of stream insect communities coloniz- Naturalist 115: 25–41. ing enclosures? Canadian Journal of Zoology 63: Wallace, J. B. & R. W. Merritt, 1980. Filter-feeding ecology of 1519–1530. aquatic insects. Annual Review of Entomology 25: Peckarsky, B. L., 1986. Colonization of natural substrates by 103–132. stream benthos. Canadian Journal of Fisheries and Aquatic Wallace, J. B. & J. R. Webster, 1996. The role of macroinver- Sciences 43: 700–709. tebrates in stream ecosystem function. Annual Review of Peckarsky, B. L. & S. I. Dodson, 1980. Do stonefly predators Entomology 41: 115–139. influence benthic distributions in streams? Ecology 61: Wallace, J. B., S. L. Eggert, J. L. Meyer & J. R. Webster, 1999. 1275–1282. Effects of resource limitation on a detrital-based ecosys- Quinn, J. M., A. B. Cooper, M. J. Stroud & G. P. Burrell, 1997. tem. Ecological Monographs 69: 409–442. Shade effects on stream periphyton and invertebrates: an Wang, L., D. M. Robertson & P. J. Garrison, 2007. Linkages experiment in streamside channels. New Zealand Journal between nutrients and assemblages of macroinvertebrates of Marine and Freshwater Research 31: 665–683. and fish in wadeable streams: implication to nutrient cri- Richards, C. & G. W. Minshall, 1988. The influence of teria development. Environmental Management 39: periphyton abundance on Baetis bicaudatus distribution 194–212. and colonization in a small stream. Journal of the North Williams, D. D., 1980. Temporal patterns in recolonization of American Benthological Society 7: 77–86. stream benthos. Archiv fur Hydrobiologie 90: 56–74. Richardson, J. S., 1991. Seasonal food limitation of detritivores Williams, D. D. & H. B. N. Hynes, 1977. Benthic community in a montane stream: an experimental test. Ecology 72: development in a new stream. Canadian Journal of Zoology 873–887. 55: 1071–1076. Robinson, C. T. & G. W. Minshall, 1986. Effects of disturbance Wise, D. H. & M. C. Molles, 1979. Colonization of artificial frequency on stream benthic community structure in rela- substrates by stream insects: influence of substrate size and tion to canopy cover and season. Journal of the North diversity. Hydrobiologia 65: 69–74. American Benthological Society 5: 237–248. Yuan, L. L., 2010. Estimating the effects of excess nutrients on Robinson, C. T., G. W. Minshall & S. R. Rushforth, 1990. stream invertebrates from observational data. Ecological Seasonal colonization dynamics of macroinvertebrates in Applications 20: 110–125.

123